section image

SafeRoute AI: Smart logistics solution for real-time truck tracking

SafeRoute AI: Smart logistics solution  for real-time truck tracking
System for Analyzing Data from Trucks Obtained Using GPS, CAN, and Driver Behavior Sensors
Key Results
100%
Processing of telematics events in a standardized format
Reduction
in route and speed violations
Real-time
reporting on fuel consumption, downtime, and violations
Client
We were approached by a client, a large transport and logistics company that manages an extensive fleet of modern trucks, including those involved in international freight transportation. Since the company's priority was to improve the efficiency of logistics routes, reduce vehicle wear and tear, and improve transportation safety, it was decided to implement a custom GPS and accelerometer-based fleet monitoring solution for deep analytics of telematics data collected from various sensors on board each vehicle. A key requirement for this custom truck telematics data analysis solution was to eliminate expensive streaming infrastructure and utilize existing tools, including a solution for tabular data analysis.
Objective
The WEZOM team was tasked with developing a commercial vehicle GPS tracking and monitoring platform that could automatically process data from Teltonika FMB920 trackers, record key events on the route (idle times, route deviations, speeding, etc.), and identify dangerous driver behavior on the road, including aggressive driving. We also needed to develop an interface for truck telematics integration with ERP/TMS/WMS, eliminate the need to connect to a complex DBMS, and provide the ability to work with flat files in Excel. At the same time, real-time streaming analytics in this driving behavior monitoring system for fleets was not required.
Location:
UkraineUkraine
Development time:
3 month
Cooperation period:
2 years
Project Team
DevOps
AI manager
AI Ops Manager
Generative AI Product Manager
Work Approach
Data collection and standardization
Route event modeling
Data processing and flows

Data collection and standardization

This vehicle sensor data visualization and reporting tool is based on Teltonika FMB920 telematics devices installed on all customer trucks. They transmit data via GPS, accelerometer, and CAN bus, and their update frequency reaches 10 Hz. All messages in this CAN bus data analysis software for trucks are transmitted via Teltonika's proprietary protocol over TCP/UDP and are converted to a unified JSON format on the client side.

Route event modeling

The algorithms record vehicle behavior based on current coordinates, speed, and accelerometer data. They determine the route deviation (if it diverges from the GPX trajectory), idle time (if speed = 0 during a specified interval), speeding (either by CAN parameters or by calculating coordinates/time), and aggressive driving (with sharp acceleration/braking calculated using the accelerometer). The fleet data analysis system also required ADAS scenarios based on processing events such as collision.warning, lane.departure, and distance.event, algorithms aimed at generating reports, and a number of integrations.

Data processing and flows

Data processing in this route deviation and geofence alert system occurs in batch mode: JSON files are exported to Excel tables without using a DBMS. This enables flexible analysis even in limited computing environments. Additionally, we implemented integrations with the client's corporate systems via REST API and an MQTT broker, which allowed us to synchronize this custom truck fleet sensor data analytics system with external TMS/ERP systems, logistics exchanges, and client portals.

Technical Architecture
Teltonika FMB920
Teltonika proprietary protocol (TCP/UDP
JSON
Excel (flat files)
Python processing (batch mode)
REST API
MQTT broker

Teltonika FMB920

Teltonika FMB920 – universal tools for real-time GPS tracking for trucks with a CAN interface and accelerometer, which enable receiving extended telematics data (speed, acceleration, fuel consumption, engine status, etc.) with a high frequency of up to 10 Hz. The client chose them for their reliability, compact form factor, and compatibility with international FMS standards.

Teltonika proprietary protocol (TCP/UDP

Teltonika proprietary protocol (TCP/UDP) – this secure truck data transmission protocol provides reliable transmission of large amounts of data with minimal delays; its advantage is the ability to be used to deliver messages from the tracker to the central system without intermediate gateways.

JSON

JSON – a unified message format, which was chosen as a data exchange standard, providing structure, readability, and compatibility with various analysis and integration tools; in this fuel consumption analysis tool for logistics, it ensures processing telematics regardless of the device model.

Excel (flat files)

Excel (flat files) – are used in this route deviation detection tool as the main solution for storing and analyzing data, as they do not require installation and maintenance of a DBMS. This tech stack component also ensures that analytics is accessible even to users without a technical background.

Python processing (batch mode)

Python processing (batch mode) – was used in this driver behavior analytics solution for batch processing of JSON messages and converting them to Excel; we chose it for its wide ecosystem for data processing and flexibility in setting up pipelines.

REST API

REST API – provides fleet telematics integration with ERP, TMS, and a number of client platforms, with access to route events, statuses, and reports. It also provides easy integration via standard HTTP requests.

MQTT broker

MQTT broker – implements a subscription to events in real time (e.g., route deviation, speeding, etc.). Our team chose it for GPS and accelerometer data analysis in transportation for its extensive scalability, compatibility with IoT devices, and low network load.

Results Achieved
Centralized data analysis
Route control and behavioral analysis
Integrations for full synchronization of data streams

Centralized data analysis

We were able to effectively combine data from hundreds of trackers in a single JSON format in this vehicle telematics analytics solution, thereby eliminating the data chaos associated with differences between devices. Thanks to the flat file format, analytics in this fuel efficiency monitoring software can be performed in Excel, which, in turn, reduces server costs and simplifies data access for managers.

Route control and behavioral analysis

The truck sensor data platform automatically detects deviations from the route, as well as records downtime, speeding violations, and aggressive driving. All events can be viewed both in reports and via API. Support for ADAS scenarios strengthened control over traffic safety and gave the client an effective telematics data visualization tool for preventing potential emergency situations.

Integrations for full synchronization of data streams

We provided a full-fledged REST API and MQTT broker in this CAN bus vehicle diagnostics solution, allowing the client to connect TMSes, WMSes, or ERPs via webhook or broker connectors. Our team has also made it possible to connect this harsh driving detection system to third-party logistics platforms or client portals – this is done through the same REST/MQTT API and configured intermediate connectors.

We use cookies to improve your experience on our website. You can find out more in our policy.